Summary of Global GPU Chip Development Companies

Work hard for a day, and you can enjoy a peaceful night,Work hard for a lifetime, and you can achieve a long-lasting happiness.

Introduction:

To assist everyone in job searching or switching careers, I have compiled thiscomplete list of global GPU chip development companies (including discrete graphics cards, mobile SoC integration, GPGPU, IP licensing, and automotive-grade chips).

First Tier: Global GPU Dominators

This tier monopolizes the global high-performance PC, workstation, and over 90% of the data center AI training market, serving as a benchmark for verification methodologies and architectural design.

1. NVIDIA (NVIDIA) 🇺🇸

  • Positioning: The absolute ruler in the AI era, the “gatekeeper” of the software and hardware ecosystem.

  • Core Products:

    • Data Center: Blackwell architecture (B100/B200/GB200), Hopper architecture (H100/H200).

    • Consumer: GeForce RTX 50 series (Blackwell), RTX 40 series.

    • Embedded/Automotive: Jetson Thor, DRIVE Orin.

  • Brief Review: The CUDA ecosystem is its deepest moat, and the architecture benchmark most familiar to verification engineers.

2. AMD (Advanced Micro Devices) 🇺🇸

  • Positioning: The only challenger with a complete high-performance CPU+GPU supply chain.

  • Core Products:

    • Data Center: Instinct MI355X / MI400 series (CDNA architecture).

    • Consumer: Radeon RX 9000 series (RDNA architecture).

  • Brief Review: With the ROCm open ecosystem and cost-effectiveness, it is aggressively capturing market share in supercomputing and data centers.

3. Intel (Intel) 🇺🇸

  • Positioning: Leveraging its PC market advantage, it is counterattacking with discrete graphics and AI chips.

  • Core Products:

    • Consumer: Arc (Acer) B series (Battlemage architecture).

    • AI Acceleration: Gaudi 3 (optimized for large model training).

    • Integrated Graphics: Xe2 / Xe3 architecture (integrated into CPUs like Lunar Lake).

Second Tier: Domestic GPU New Forces

This is the most active segment globally, divided into two camps: “Graphics Rendering” and “GPGPU Computing Power”.

Camp A: Focused on Full-Function Graphics Rendering

Goal: To replace desktop graphics cards, adapt to domestic operating systems, and solve the “no graphics with a U” problem.

4. Moore Threads

  • Products: MTT S80/S30 (desktop), MTT S4000 (server).

  • Brief Review: Focuses on “full-function GPU” and MUSA unified architecture, being the most proactive manufacturer in supporting DirectX and gaming among domestic cards.

5. Jingjia Micro

  • Products: JM9 series, JM9271.

  • Brief Review: Originating from a military background, a veteran in domestic GPUs, with a high market share in trust innovation and special fields, now expanding into civilian use.

6. Innosilicon

  • Products: Fantasy 1/2.

  • Brief Review: Possesses strong IP design capabilities, focusing on multi-path cloud rendering, cloud gaming, and desktop office applications.

7. Loongson

  • Products: 9A1000, LG series discrete graphics cards.

  • Brief Review: Adheres to self-developed instruction sets, building a completely autonomous computer system in conjunction with Loongson CPUs.

8. Glenfly

  • Products: Arise GT10C0.

  • Brief Review: Leveraging VIA’s technology background, mainly collaborating with Zhaoxin x86 CPUs for government and enterprise office displays.

9. Chipintelli

  • Products: GenBu series, GB2062.

  • Brief Review: Focused on high-performance graphics rendering for domestic computers, emphasizing low power consumption and operating system compatibility.

10. Superdisplay

  • Products: Tianyuan series GPU cores.

  • Brief Review: Focused on embedded GPU IP and customized chips, applied in industrial control instruments, etc.

11. Hangjin Tech

  • Products: Military/special graphics processing chips.

  • Brief Review: Under Changsha Shaoguang, focusing on high reliability and military-grade reinforced graphics processing chips.

12. Xiangdixian

  • Products: Tianjun series (though there have been recent rumors of restructuring, it remains an important historical participant in the industry).

  • Brief Review: Previously launched high-performance general-purpose GPUs, targeting the mid-to-high-end desktop market.

Camp B: Focused on GPGPU and AI Computing Power

Goal: To compete with NVIDIA A100/H100, building a domestic intelligent computing foundation.

13. Biren Technology

  • Products: BR100 / BR200 series.

  • Brief Review: Adopts Chiplet design, with extremely strong single-card computing power, one of the representatives of high-end general-purpose GPUs in China.

14. MetaX

  • Products: Xiyun C500 (training), Xisi N100 (inference), Xicai (rendering).

  • Brief Review: The core team originates from AMD, possessing a complete software stack, with a product line covering training, inference, and graphics.

15. Iluvatar CoreX

  • Products: Tianhai (BI) series, Zhikai (MR) series.

  • Brief Review: The first domestic GPGPU company to achieve mass production, emphasizing smooth migration and compatibility with the CUDA ecosystem.

16. Enflame

  • Products: Suisi (CloudBlazer) T20/S20.

  • Brief Review: Invested by Tencent, focusing on cloud AI computing power, with experience in large-scale cluster deployment (Note: previously misreported as TuSen, it is an independent AI chip company).

17. Cambricon

  • Products: Siyuan (MLU) 590/370 series.

  • Brief Review: “The first stock of AI chips”, although the technical route leans towards NPU/ASIC, it occupies an important share in domestic intelligent computing centers.

18. Denglin Tech

  • Products: Goldwasser (GPU+) series.

  • Brief Review: Adopts a software-defined heterogeneous architecture, compatible with CUDA, excelling in edge computing and security fields.

19. Vastai

  • Products: VA / SG series.

  • Brief Review: Entered from video encoding/decoding (VPU), possessing strong video AI processing and cloud gaming rendering capabilities.

Third Tier: Mobile SoC and GPU IP Licensing

The graphics cores in your mobile phones, tablets, and automotive cockpits mostly come from here.

Chip Design (Fabless)

20. Apple (Apple) 🇺🇸

  • Products: M4 / A18 Pro series built-in GPU.

  • Brief Review: The leader in mobile energy efficiency and single-core performance, with a fully closed Metal ecosystem.

21. Qualcomm (Qualcomm) 🇺🇸

  • Products: Adreno GPU (Snapdragon 8 Elite).

  • Brief Review: The strongest GPU in the Android camp, now extending its reach into the PC market through Snapdragon X Elite.

22. HiSilicon (Huawei) 🇨🇳

  • Products: Maleoon (Maliang) 910/920 GPU.

  • Brief Review: Integrated into Kirin 9000S/9100 series, self-developed architecture breaking the blockade, performance approaching mainstream levels.

23. Broadcom (Broadcom) 🇺🇸

  • Products: VideoCore series.

  • Brief Review: The “heart” of Raspberry Pi, widely used in set-top boxes and embedded fields.

IP Licensing Suppliers (IP Licensor)

24. Arm Holdings 🇬🇧

  • Products: Mali / Immortalis-G925.

  • Brief Review: The number one in mobile GPU IP market share, a key partner for MediaTek and Samsung.

25. Imagination Technologies 🇬🇧 (Chinese background)

  • Products: PowerVR / IMG DXT series.

  • Brief Review: The pioneer of mobile GPUs, currently leading in ray tracing IP and automotive electronic safety fields.

26. VeriSilicon 🇨🇳

  • Products: Vivante GPU IP.

  • Brief Review: The largest semiconductor IP supplier in China, with a high market share in automotive cockpit and IoT display fields.

Fourth Tier: Automotive Electronics and Automotive-Grade Chips

With the development of smart cars, automotive-grade GPUs have become a new battleground.

27. Horizon Robotics 🇨🇳

  • Products: Journey 6 series.

  • Brief Review: The leader in domestic intelligent driving computing solutions, with BPU architecture optimized for intelligent driving.

28. Black Sesame 🇨🇳

  • Products: Huashan/Wudang series.

  • Brief Review: A cross-domain computing platform covering autonomous driving and intelligent cockpits.

29. NXP (NXP) 🇳🇱 / TI (Texas Instruments) 🇺🇸 / Renesas (Renesas) 🇯🇵

  • Products: i.MX series / TDA4 series / R-Car series.

  • Brief Review: Traditional automotive electronics giants, mainly applied in central entertainment systems, dashboards, and driver assistance.

Fifth Tier: Emerging AI Architectures

Taking an unconventional path, challenging NVIDIA with non-traditional GPU architectures.

30. Tenstorrent 🇨🇦

  • Products: Grayskull / Wormhole.

  • Brief Review: Led by chip guru Jim Keller, an AI processor based on RISC-V.

31. Graphcore 🇬🇧

  • Products: IPU (Intelligent Processing Unit).

  • Brief Review: A graph computing architecture designed for machine intelligence (acquired by SoftBank in 2024).

32. Cerebras 🇺🇸

  • Products: WSE-3 (wafer-scale engine).

  • Brief Review: A wafer is a chip, specializing in ultra-large model training, with a strong visual impact.

33. Groq 🇺🇸

  • Products: LPU (Language Processing Unit).

  • Brief Review: Challenging GPUs with unprecedented inference speed (Token/s), designed specifically for LLM inference.

Global GPU Landscape:

  • High-end: Still dominated by the three-way competition of NV / AMD / Intel.

  • Domestic: Has formed a “3+N” army, with 2025 being the year for large-scale deployment of domestic GPUs.

  • Trends: The boundary between graphics and AI is blurring, with automobiles becoming the next computing center.

(Data statistics are as of November 2025; if there are any omissions, please feel free to supplement and correct in the comments section!)

*Disclaimer: This article is original by the author. The content reflects the author’s personal views, and the reprint by LuKe Verification is only to convey a different perspective, not representing LuKe Verification’s endorsement or support of this view. If there are any objections, please contact LuKe Verification.

Related posts

  1. Considerations for PCB and PCBA Design During Drop Testing
  2. Unit 3: Size Inspection of PCB Connection Pads
  3. Why is the Training of Large AI Models Dominated by the Python Ecosystem?
  4. University Turns into Chip Factory? Fuyao Technology University Makes Waves with Bold Move
  5. Regret Not Increasing Investment in Sino-Korean Semiconductors!
  6. How AI is Disrupting Chip Design
  7. How to Choose Between NSA and SA Networking Modes? Operators Indicate…
  8. A Batch of University Major Adjustments! Popular Fields like Intelligence and Robotics ‘Launched’
  9. Six Emerging Companies in the Third Generation Semiconductor Sector (Selected List)
  10. Review of PCB Market Trends from May 6 to September 5
  11. Elon Musk: The AI5 Chip Will Reshape the Era of Edge AI; Its Relevance to the Latest Trillion-Dollar Compensation Bet!
  12. The Semiconductor Testing Equipment Breakthrough: Domestic Substitution is Timely
  13. The AI Hardware Bubble: The Harsh Truth Behind the Exhibition’s Excitement, 99% of Products Are Just Transitional
  14. JHM: Visual Immunosensor Method with Double T Line for On-Site Sensitive Detection of Fipronil Pesticide in Water, Soil, and Honeysuckle
  15. Geely Constellation Completes Phase One Network Deployment, Ushering in a New Era of “Integrated Space and Earth” for Chinese Automotive Industry!
  16. China’s First! Geely Constellation Completes Phase One Networking, 64 Satellites Weave a Global “Sky Net”
  17. Domestic 5nm Self-Developed GPU Completes Tape-Out Verification! Xiangdi’s Next-Generation Fuxi Architecture Officially Announced
  18. Storage Chips and MCU Chips: The Rise of Eight Leading Companies
  19. The Significant Impact of the Sequence in PCBA Processing: Soldering Before Separation

Leave a Comment